Gene mapping for alcohol use disorder (AUD) has been challenging due to phenotype heterogeneity and complicated genetic architecture. Traditional genome-wide association study (GWAS) has focused on common single nucleotide variant (SNV) in case-control sample. To address these limitations, we propose to detect genes for AUD by applying the exome-focused analysis in a longitudinal cohort of patients with and without HIV infection. We will conduct a longitudinal GWAS (LGWAS) in 2,470 veterans from the Veterans Aging Cohort Study (VACS). We have identified 3 AUD trajectories in this population (No AUD; Moderate AUD, and Severe AUD). We will apply a cost-effective and powerful approach to conduct a longitudinal GWAS for gene detection. We will use the Illumina HumanCoreExome Beadchip that compasses 560K low-frequency putative functional SNVs and common tag SNVs. We have successfully genotyped 1,000 out of 2,470 DNA samples using this array. Additionally, we have developed a novel statistical method to analyze gene-based LGWAS for this project. In this study, we aim 1) to identify genetic risks for AUD trajectory by using a longitudinal cohort. We hypothesize that low-frequency variant in a gene collectively contribute to associate to AUD and can be detected by LGWAS in a moderate sample size; 2) to perform gene set enrichment analysis (GSEA) to identify biological pathways related to AUD trajectories. GSEA approach is able to detect marginal significant genetic effects. In addition, we will examine the interaction of genes and HIV status in AUD progression. We hypothesize that HIV exposure manifests marginal genetic variant effects on AUD and a subset of novel genes will be revealed through testing the interaction of genes and HIV status. Long-term objectives: Our goal is to identify risk individuals for AUD progression. We will evaluate neurobiological mechanisms of identified variants and pathways from this study. We hope to provide biological evidence for future medication development and individualized care.